To address the widespread challenges of insufficient classroom participation, difficulty maintaining learning interest, and inaccurate learning outcome assessment in existing teaching models, this study introduces a hybrid and dual-teacher classroom model. Based on big data and online platforms, this study constructs a music teaching resource system that integrates online micro-lessons, real-time interaction, virtual choirs, collaborative composition, and learning behavior tracking to enhance the openness and personalization of teaching. Through comparative experiments, the study focuses on changes in student knowledge acquisition, musical skills, learning interest, and classroom engagement. Results show that the experimental group significantly outperforms the control group in terms of improvement in knowledge test scores, rhythm and pitch performance, task completion rate, and scores on the learning motivation questionnaire. Differences are particularly prominent in interaction frequency and the expressiveness of the work. The study shows that Group A demonstrates significant improvement in all three dimensions: interest increases from 3.1 to 4.2; autonomous learning motivation increases from 3.0 to 4.1; emotional engagement increases from 3.2 to 4.3, demonstrating strong positive effects. Experimental Group B shows the greatest improvement, with interest increasing from 3.0 to 4.5, autonomous learning motivation from 2.9 to 4.4, and emotional engagement from 3.1 to 4.6, indicating the most significant improvements in learning attitudes and emotions.